Get Proababilities from Choice(coordination) Sub-Model

[non87]

Hello everyone,

it's been a while!

I am trying to interpret the results I got from the choice-coordination models I estimated. I am trying to calculate an average marginal effect for the main coefficients in my models. The issue I am facing is that I can work out the math, but it is not easy to implement it in R using the Goldfish package -- at least, it is not clear to me.

Substantially, an actor evaluates the convenience of a given change in the network based on the utility she would get by changing the network. So, she evaluates the utility of the network supposing it will change in a certain way. While the linear objective function of an actor is easy enough to calculate given the statistics and the coefficient, I do not know how to calculate the statistics and I would guess goldfish must do it at some point during estimation.

To be very specific, I would like to calculate equations (1) and (2) from Statfield et al., Dynamic Network Actor Models: Investigating Coordination Ties Through Time, Sociological Methodology, 2017 , but I am stuck because I not know how to conveniently calculate the s_m(*) in equation (1) from my data.

Any help appreciated, thanks!